AMD GPU 프로그래밍 입문에서 전문가까지 (1부) - Composable Kernel(CK)의 TensorDescriptor

auto tensor_desc = make_naive_tensor_descriptor(make_tuple(M, K), make_tuple(K, 1));

auto transformed_tensor_desc = transform_tensor_descriptor(
    tensor_desc,
    make_tuple(unmerge, passthrough),
    make_tuple(Sequence<0>{}, Sequence<1>{}), // Lower dimension ids
    make_tuple(Sequence<0, 1>{}, Sequence<2>{}) // Upper dimension ids
);

static_for<0, 4, 1>{}([&](auto i){
    a(Number<i>{}) = buf.Get<d4_t>(
        tensor_desc.CalculateOffset(Tuple{x + i, y}), true);
});

// In-register transpose
static_for<1, 4, 1>{}([&](auto i){
    static_for<0, i, 1>{}([&](auto j){
        auto tmp = b(Number<i * 4 + j>{});
        b(Number<i * 4 + j>{}) = b(Number<j * 4 + i>{});
        b(Number<j * 4 + i>{}) = tmp;
    });
});